The Generations Project: An Engine for Both Patient Care and Discovery

April 30, 2021

ID6546

To CiteDCA Citation Guide

00:00My name is Sam Powell.
00:02I'm an MD, PhD student and currently
00:05in my final year of the PhD in the
00:09lab of Kristen Byrne and it's my
00:11great pleasure to introduce our next
00:14speaker who is Doctor Mike Murray.
00:17Doctor Murray joined Yale in
00:192018 as the Director of Clinical
00:21Operations in the newly formed
00:24Yale Center of Genomic Health.
00:26Prior to coming to Yale,
00:28he was the Director of Genomic Medicine.
00:31At Geisinger Health System,
00:33where he led the design and implementation
00:35of the Genome First Program in 2015,
00:38he is a physician and his board certified
00:41in internal medicine and medical genetics,
00:44and is the principle investigator
00:46on the Yale Generations Project.
00:48So with that I will turn over the
00:51virtual floor to Doctor Murray.
00:59I think you're still muted,
01:00muted Doctor Mary.
01:03Thank you, thanks Sam and thanks
01:05to all the organizers and everyone
01:07that's worked on this. I have the.
01:11The honor and the opportunity to tell you
01:13about the Generations project and it is
01:16the work of dozens and dozens of people
01:19around the health system and the school.
01:22So I will tell you about it in 15 minutes
01:26and hopefully have a little bit of
01:28time for for some questions at the end.
01:32So when we describe the generations
01:34project we talked about it as a
01:37large sequence cohort project aiming
01:39to get at least 100,000 volunteers
01:41over the first five year period.
01:44And to be an engine for both 21st century
01:47Healthcare and Discovery Research,
01:49I have no see any conflicts of interest
01:52for this event when we talk about the
01:56deliverables from the Generations project.
01:59There's four main areas.
02:01First is data for research,
02:03so DNA datasets linked to HR phenotypes
02:05are available for researchers.
02:07They can be either identified or
02:10deidentified depending on the IRB
02:12approval that the researchers have.
02:14We have banked biospecimens.
02:16Right now, it's just germline DNA,
02:19but over time we anticipate adding
02:21other types of bio specimens.
02:24There are participants.
02:25All the volunteers are consented
02:27for Recontact,
02:28so there potentially.
02:29Participants for new studies,
02:31including clinical trials,
02:33surveys the callbacks for deep phenotyping
02:36biospecimens another and the last
02:38deliverable is precision medicine,
02:39so everyone that volunteers into the
02:42project receives back clinical results
02:44into their electronic health record.
02:47And I'll describe how that works
02:50to you in the slides ahead.
02:54So we spent about a year and a half
02:56setting up the workflows and building
02:59this and and this workshop as well.
03:01Time because we're now really at the
03:03point where we can open this up broadly,
03:06an invite health care providers and
03:08researchers around Yale to think about
03:11this unique infrastructure to build on
03:13it and to collaborate with this project.
03:16So what are we talking about when we
03:18talk about 21st century healthcare?
03:20We describe it like this.
03:22A patient who volunteers has
03:24a comprehensive data set,
03:25genomic data set that's created
03:27and then held securely within the
03:29health system and over a lifetime.
03:31We anticipate that that data set can
03:33be used to generate test results.
03:35There's two types of test results that
03:38we anticipate and are planning on.
03:40The first are screening results,
03:42so looking at an individual's data
03:44and looking for otherwise invisible
03:46risks for heart disease or cancer or
03:48other diseases that are there and are.
03:50Present,
03:51but they and their health care
03:53provider may not know about the second
03:56category are clinically indicated.
03:57Diagnostic test results.
03:58So when a patient presents with a medical
04:01problem and it's part of generations,
04:04we anticipated overtime,
04:05more and more opportunities will
04:07be available to order a clinical
04:09test through generations.
04:10So the test results that are ordered
04:12in this way will get delivered
04:14into the EHR without a new patient
04:17sample without a repeat blood drawn
04:19without any additional sequencing
04:21or genotyping. And then the
04:24results are managed by clinical
04:26experts within the health system.
04:28Highlighted in red are the
04:30two areas that really are new
04:32within the generations project,
04:34and as far as I know are not being
04:37offered or anticipated anytime
04:39soon within similar sequence
04:41cohorts at other institutions.
04:44So more broadly, about generations,
04:46as mentioned earlier,
04:47this is a collaboration between
04:50the health system in the school,
04:53anticipating at least 100,000
04:55volunteers over a five year period.
04:57The DNA samples undergo both zoman sniper,
05:01a data generation.
05:02There are no exclusion criteria,
05:05any age, any health status can
05:08volunteer into this project,
05:09and the data gets linked to.
05:13To the individuals electronic health record.
05:17So the consenting is.
05:20Is a little complicated
05:22because of that integration,
05:23so it's different than maybe some other
05:26consenting that you're familiar with and
05:28the IT Group has really made this happen,
05:30and we're grateful when someone wants
05:32to volunteer for four generations if
05:34they have never been seen at the Yale
05:37New Haven health System and don't have
05:39an electronic health record or consent team,
05:42is able to generate an electronic
05:44health record during the process.
05:46So in many registration,
05:47all volunteers then sign a consent.
05:49It gets direct, directly linked.
05:51Into their epic record and that
05:54consenting triggers a physician
05:56order for DNA test within Epic.
05:58This is a clinical order.
06:00This is not a research request and
06:03that DNA sample goes through the
06:06normal clinical workflows through
06:08the laboratories and off to the DNA
06:12diagnostic lab in the West Campus for
06:15for genomic sequencing and genotyping.
06:17The DNA can be acquired from
06:21either blood or saliva.
06:23The deliverables we talk about it
06:26as all the results that we plan on
06:29giving back are actionable results.
06:32They have clinical utility,
06:34meaning that of provider and the
06:36volunteer can use that to make clinical
06:39decisions and the examples of broad
06:42categories are monogenic screening,
06:44pharmaco,
06:45genomics and diagnostic testing.
06:47Following the screening,
06:48we give back the positive results
06:51through the genomic health team,
06:53and there's three steps that
06:54go with each return of result.
06:57That is,
06:58the first step is education and counseling,
07:00essentially genetic
07:01counseling about the result,
07:03what it means and what it doesn't mean,
07:06and then family cascade testing
07:08is offered so that if someone has
07:11a risk result that we uncover,
07:13then their siblings,
07:14children and parents are offered
07:17participation so that they can
07:19find out if they have the same.
07:21Same risk and then referral to
07:25clinical experts for management
07:28and evaluation overtime.
07:31There's no cost for participation
07:33to the to the volunteer,
07:34the DNA analysis,
07:36and the Genomic health visit are
07:38part of being in the project.
07:41This is data that we pulled
07:43recently on the approximately
07:443500 patients that are currently
07:46enrolled in the Generations project.
07:49On the top panel you can see that
07:51the cohort currently is made up
07:54of primarily of adults age 2280.
07:56However,
07:57there are children and there
07:59are older individuals that have
08:01been consented into the project,
08:03including an individual over the
08:05age of 100 in the middle and lower
08:08panel or pie charts showing the.
08:10Race and ethnicity of the participants.
08:13It's important to note that this
08:15race and ethnicity is drawn directly
08:18from the electronic health record,
08:20and so both both of those pie charts
08:23have relatively large slices that are
08:26unknown because the electronic health
08:28record doesn't doesn't require this.
08:31This category of information currently,
08:33so we're working on strategies
08:35to fill those gaps overtime.
08:40Her two other pie charts, the one on the top,
08:44talks about those that had an electronic
08:47health record versus those that had one
08:50generated in order to become volunteers
08:53and generations 17% in the dark blue.
08:55There are those that had a new epic record
08:59created for participation in generations,
09:02and then the bottom pie chart
09:05shows the approximately 75% of
09:07individuals who were new registrants.
09:09Via the process,
09:10who then later had a clinical
09:13encounter of some sort following it.
09:15So though we haven't been specifically
09:18targeting individuals who are not
09:20Yellow Haven hospital patients,
09:21they are coming to us and they are
09:24not only signing up for generations,
09:27but they're then getting some elements
09:29of their care within the health system.
09:34Shown here on the left is our
09:37monogenic risk screening table.
09:38There are 10 genes associated with
09:41monogenic risk for cancer or heart disease,
09:4310 out of the 20,000 that each of us have.
09:47So I jokingly say that this version
09:49one will go through many versions,
09:52be before we eventually get to the 20,000
09:54total that will ultimately give back,
09:57and then on the right is the Pharmaco
09:59genomics panel that we currently offer back.
10:02There are six genes associated
10:04with the medications that you
10:06can see on the far right.
10:08Column that are given
10:10back into the epic record,
10:11and then there's clinical decision
10:14support driven off of that.
10:16A lot of folks ask what's the
10:18percentage of people that get
10:20back results based on this?
10:22We're currently giving back about
10:24as you can see there 1.4% of
10:26volunteers are getting a result
10:28back on this short list of 10 genes
10:30and conditions for monogenic risk,
10:32that number will go up later this
10:35year when we go to our version two,
10:37which will have 44 genes in the
10:40panel and this is consistent with
10:42with other groups as far as the
10:44percentage is receiving back results.
10:46On the pharmacogenomics,
10:48there are now over 2000 individuals
10:50that have pharmaco genomic
10:51results in their epic record.
10:53As a result of participation in generations.
10:57Some of you will be familiar
11:00with this published data.
11:01This came out of work that I was involved in,
11:05and Geisinger Health System in
11:06Pennsylvania where we did a
11:08similar sequence cohort project.
11:10This was on the 1st 50,000 individuals
11:12who participate in that project.
11:14We went back and we looked at
11:16their data for those individuals
11:18that had changes in their BRCA,
11:21one or two gene that would be
11:23expected to confer risk for cancer.
11:26We found that about one in 190
11:28individuals in this population.
11:30Had one of those risks and 80% of
11:33those individuals had no idea that
11:35they had this risk the first time
11:38that they found out was through
11:40participation in this sort of a sequence
11:44project project with with population
11:46screening for this particular risk.
11:48So this is the DNA based screening
11:51that we anticipate and plan to
11:53roll out to all participants.
11:55It offers the chance to uncover
11:58invisible risks such as this and others.
12:01And and will go to.
12:03As I mentioned earlier,
12:0444 genes and conditions associated
12:07with risk can expand that overtime.
12:10But I promised that I'd also talk
12:12about diagnostic use of the data and
12:14so I want to go into the first use
12:17case that we've launched around this.
12:19And this builds off of paper that
12:21showed up in the New England
12:24Journal of Medicine written by
12:26the team at Columbia in New York.
12:29And they offered xom sequencing
12:31on a research basis to individuals
12:33with kidney disease and reported
12:35out the diagnostic utility of
12:38screening those individuals.
12:39The renal team here.
12:42Came to us in we discussed the fact
12:45that they can't easily order this kind
12:47of testing that they love to for the
12:51patients that are listed for transplant.
12:53There's all kinds of hurdles to
12:55getting this kind of testing ordered
12:57and resulted within healthcare.
12:59Currently zero.
13:00I saw this as an opportunity to
13:02to explore this opportunity for
13:04down the middle of the slide.
13:07Here you can see in blue the
13:09nephrology just wanting to simply
13:11place a clinical order.
13:13In epic for tests that had the potential
13:16to offer back positive results,
13:1829 percent of their transplant patients
13:21that would influence their care,
13:23and we had the aspiration of using the
13:26generations data for this kind of use.
13:29So to do that we had to move this
13:32down to the bottom and up top.
13:35Here offer the patients participation
13:37in generations as they got identified
13:40as renal transplant candidates.
13:41So we started this.
13:43Back in the fall together with the
13:47YCCI based, consenting for generations,
13:50we now have about 65 individuals.
13:54Who are on the transplant list and
13:57have consented into generations
13:58and their receiving the standard
14:01used to the data, the screening,
14:03and the Pharmaco Genomics as mentioned.
14:06And then just this week,
14:08the epic team worked hard and stood
14:10up a Go live order that are not
14:13for Ologist can now place an order
14:16for what we're calling the renal
14:19gene screen
14:20panel. 27 genes associated with
14:22renal failure that they'll be able
14:25to offer going forward and hopefully
14:28within the coming months will
14:30have some data to share about the
14:33implementation of this strategy.
14:37Other diagnostic uses of data, of course.
14:40I'm sure that individuals can
14:42think of many different ways,
14:44and so you know we are open to
14:47hearing your suggestions and
14:49working with you to stand these U.
14:53The next case that we expect to
14:55stand up is in the smilow breast
14:58cancer program where we plan to offer
15:01enrollment in generations early in
15:04the diagnosis and then deliver back.
15:07Results of germline diagnosis to
15:08those individuals who have been
15:10diagnosed with cancer,
15:11but there's many more potential use
15:14cases and I've already talked to
15:16a number of people about others.
15:20Regarding the research use of generations
15:22data, you know some of you have heard
15:25me describe generations as a living
15:27laboratory because of the generosity
15:28of the volunteers they have made
15:31their data available for research,
15:33and they've also made themselves
15:35available for callbacks.
15:37And so there's three categories of data
15:39or callbacks that we're anticipating.
15:42The first is healthy controls,
15:44so researchers that have a group of
15:47patients with the disease or condition
15:49that are looking for age or gender or
15:53other criteria matched healthy controls.
15:55We anticipate having those controls
15:57available for either sharing of data sets
16:01or for calling back individuals for studies.
16:03As you can imagine,
16:05there also be genetically defined.
16:07Cohorts,
16:08so relatively rare snips that individuals
16:10want to study patients that have them.
16:13We can find them within that this cohort,
16:16and we've already done that together
16:19with somebody in David Hafler's Group,
16:21and then there's clinically defined cohorts.
16:24Individuals that have a
16:26medical issue or problem,
16:27and bringing back those individuals
16:29who are looking at those data.
16:32So we are,
16:33we stand ready to generate letters
16:35of support for grant proposals.
16:38To to work with you to give you genomic
16:41datasets that will require that you have
16:43an IRB approved protocol to receive
16:46such datasets and to do callbacks and.
16:48And we've we've stood this up
16:51in all three categories so far.
16:54The last thing that I'll mention
16:57is many of you familiar with this.
17:00The Nomad database is available
17:02online for individuals that want to
17:05go onto it and look up a gene or a
17:08variant and find out the frequency
17:10within this large data set,
17:12which which includes large cohorts from
17:15across the country and across the world.
17:17Luckily for us,
17:19muenkel like who's in our Department
17:21as a speaker in this workshop.
17:24Was one of the drivers of standing
17:26up this this database and will be
17:28working with us to create a mini
17:31version of this and internal HIPAA
17:34compliant version where individuals
17:35within Yale can go in and look up
17:38genes and variants within generations.
17:40Data set to do queries.
17:42So with that I'll end.
17:44I'll thank the many many people who
17:47have helped to get us this far.
17:49An R hopefully cheering us on to
17:51go further and to thank all the
17:54volunteers who have participated.
17:56In the last slide here,
17:58if you're within the sound of my voice,
18:00that means that you're a potential
18:02volunteer for generations also,
18:03and we invite you to join
18:05if you're interested.
18:06Here's how to reach us,
18:07including a QR code will leave this
18:10up during the Q&A so that you can
18:12scan your phone over and generate
18:14an email to either ask us a question
18:16or or reach the consenters who can
18:18get you consented into the project.
18:20And with that I'll stop.
18:22Thanks very much.